Hi everyone!
I'm Daryl, a third year mathematics and computer science student from the National University of Singapore, currently doing an exchange semester at EPFL. I also have some substantial experience with Python development, most notably as a (formerly active) core developer of the mathematical animation library Manim.
I'm most interested in the RUBI and benchmarking projects. The wiki states that the main issue is performance; would a project on this involve rewriting the code generation to target something like CPython, or are there other ways of optimizing the code in pure python? If the main thing taking time is MatchPy, would it involve optimizing MatchPy or integrating SymPy with MatchPyCPP?
For the benchmarking project, I have some experience with setting up self-hosted runners on GitHub actions for Manim, where we set up ARM-based VMs on Oracle Cloud to test that Manim runs correctly on ARM architecture. I was not directly involved in setting up the Linux environment, but I believe that it's something I could learn through the progress of the project.
I'm currently taking courses in Sobolev spaces and interactive theorem provers, so I would also be interested in projects involving these topics (generalized functions, and potentially an experimental project integrating with a proof assistant). However since the semester has not yet started, I'm not familiar with the specific uses this knowledge could have within SymPy.
Looking forward to hear your thoughts on these ideas!
Thanks,
Daryl